Does PCGS counterfeit technology use facial recognition type of software?

60 streeter

I'm just wondering as the 1801 draped bust dime I dug 7 inch's down in the ground, and a mile into the woods......i researched it and it appears to be JR2... and PCGS only has two registered. PCGS could not authenticate... they didn't declare a counterfeit but..... Now I am wondering if their counterfeit software is looking for a JR1... which all the rest registered.... and the JR 2 reverse is throwing it off?

yosclimber

They don't use software to identify die varieties.
People do it, using reference books.

may

Facial recognition software, on coins??

60 streeter

@yosclimber said:
They don't use software to identify die varieties.
People do it, using reference books.

but they do use some type counterfeit software, at some point in the grading process....

Vasanti

@Ownerofawheatiehorde said:
Facial recognition software, on coins??

It's kind of the same thing as biometrics. Look at physical parameters on coins and compare them to a database to look for similarities. I'm betting they also check submissions against prior graded coins as well.

yosclimber

@60 streeter said:

@yosclimber said:
They don't use software to identify die varieties.
People do it, using reference books.

but they do use some type counterfeit software, at some point in the grading process....

Please cite your source that proves this.

Vasanti

@yosclimber said:

@60 streeter said:

@yosclimber said:
They don't use software to identify die varieties.
People do it, using reference books.

but they do use some type counterfeit software, at some point in the grading process....

Please cite your source that proves this.

https://pcgs.com/goldshield

60 streeter

@Vasanti said:

@yosclimber said:

@60 streeter said:

@yosclimber said:
They don't use software to identify die varieties.
People do it, using reference books.

but they do use some type counterfeit software, at some point in the grading process....

Please cite your source that proves this.

https://pcgs.com/goldshield

I am wondering if this is my problem? AI is looking for a JR 1, but min is a JR2 and PCGS only has a population of 2. They may not have configured to recognize the variety.......

Morgan13

How about some images?

Vasanti

@60 streeter said:

@Vasanti said:

@yosclimber said:

@60 streeter said:

@yosclimber said:
They don't use software to identify die varieties.
People do it, using reference books.

but they do use some type counterfeit software, at some point in the grading process....

Please cite your source that proves this.

https://pcgs.com/goldshield

I am wondering if this is my problem? AI is looking for a JR 1, but min is a JR2 and PCGS only has a population of 2. They may not have configured to recognize the variety.......

I seriously doubt that decisions are made solely by software, particularly when it comes to variety attribution. Nobody even knows if they use software for variety attribution, but I don't think anyone would be surprised. Software is certainly good enough to do it as a first pass, but I imagine that it's a function of cost/benefit. If software can track a human being around a city based upon how that human walks and moves, then it can certainly see die cracks and variations in strike.

https://recfaces.com/articles/what-is-gait-recognition

yosclimber

@Vasanti said:

@yosclimber said:

@60 streeter said:

@yosclimber said:
They don't use software to identify die varieties.
People do it, using reference books.

but they do use some type counterfeit software, at some point in the grading process....

Please cite your source that proves this.

https://pcgs.com/goldshield

Not really related to original counterfeit detection.

What Gold Shield does is to check the coin against prior submitted Gold Shield coins for a match.
So it can detect a stolen and cracked out coin, IF the coin was submitted to Gold Shield level both times.

The prior version of Gold Shield was Secure, which used a 3-D surface scan for matching.
Both are a more expensive level for grading.

It can also detect a coin that is a match to a coin previously determined to be a counterfeit.
The original counterfeit detection is done by a person.

PCGS Gold Shield, 2018:
https://www.youtube.com/watch?v=7JvqB1-VVoU

PCGS Secure Plus, 2011:
https://www.youtube.com/watch?v=yVf3i5xuJLQ

yosclimber

@60 streeter said:

@Vasanti said:

@yosclimber said:

@60 streeter said:

@yosclimber said:
They don't use software to identify die varieties.
People do it, using reference books.

but they do use some type counterfeit software, at some point in the grading process....

Please cite your source that proves this.

https://pcgs.com/goldshield

I am wondering if this is my problem? AI is looking for a JR 1, but min is a JR2 and PCGS only has a population of 2. They may not have configured to recognize the variety.......

I hate to say it, but this is kind of a dumb guess based on several poor assumptions.
If you are good enough to attribute a JR-2, you should be able to think more carefully than this.

To be specific, here are some of your assumptions:
1. They use computer vision software to attribute varieties.
2. They know JR-1 but not JR-2. (This makes very little sense).
3. They attributed your coin to JR-1.

What they actually said was that they were "unable to authenticate".
This can happen to a very low grade or damaged coin like yours.
At some level of very worn down or damaged features, coins become unrecognizable
in terms of die variety and even date.

Regardless of your $200 down the drain, this is still a potentially valuable coin to many collectors.
It is just not the most ideal coin to get slabbed.

60 streeter

@Vasanti said:

@60 streeter said:

@Vasanti said:

@yosclimber said:

@60 streeter said:

@yosclimber said:
They don't use software to identify die varieties.
People do it, using reference books.

but they do use some type counterfeit software, at some point in the grading process....

Please cite your source that proves this.

https://pcgs.com/goldshield

I am wondering if this is my problem? AI is looking for a JR 1, but min is a JR2 and PCGS only has a population of 2. They may not have configured to recognize the variety.......

I seriously doubt that decisions are made solely by software, particularly when it comes to variety attribution. Nobody even knows if they use software for variety attribution, but I don't think anyone would be surprised. Software is certainly good enough to do it as a first pass, but I imagine that it's a function of cost/benefit. If software can track a human being around a city based upon how that human walks and moves, then it can certainly see die cracks and variations in strike.

https://recfaces.com/articles/what-is-gait-recognition

I am thinking PCGS has so closed it self off, and seems to be very strict in what graders can and cannot do. I have a feeling they may have to give the unable to authenticate tag on anything AI flags... Check my thread 1801 draped bust dime.. this is definitely authentic!! PCGS only has 2 registered.... I think AI is not recognizing it... Probably never entered into its data base... although I find that hard to believe... But these older silvers simply do not conform as the modern ones..... check the different strike patterns in the thread

burfle23

Is this the 4th post concerning 1 coin?

Just trying to keep up...